Cocaine use is common in inner city pregnant women, in whom 10-28% had the drug detected in city hospital surveys. Study results have been conflicting in relating cocaine use to the risk of previous spontaneous abortions and there have been no studies of concurrent cocaine use and spontaneous abortions. The objective of this proposal is to conduct a case-control study to determine whether the use of cocaine is an important, independent predictor of spontaneous abortion. The proposed study is designed to be the first to ascertain from the ED, a representative cohort of inner city women who frequently do not seek prenatal care; to accurately measure cocaine use; and to evaluate the risk of spontaneous abortion in recent cocaine users as compared to non-users while controlling for the influence of other risk factors. The study will be accomplished as follows. Pregnant women at less than 22 weeks gestation presenting to the Emergency Department (ED) at the University of Pennsylvania (HUP) will be recruited. Upon entry or during follow-up through 22 weeks gestation, we will observe which subjects spontaneously abort the pregnancy and which continue to carry it. At least 225 women experiencing an incident spontaneous abortion through 22 weeks of gestation will be classified as cases. Cases will be compared to at least 450 control pregnant women who do not experience a spontaneous abortion. Controls will comprise two separate groups to which Cases will be compared: those with vaginal bleeding i.e. with threatened spontaneous abortion and those without bleeding. Women with ectopic or molar pregnancies, twin pregnancies, or elective abortions will be excluded from analyses. Data collection will consist of a baseline structured in-person interview including questions about drug use. Baseline laboratory assessment will include urine screening for cocaine, marijuana, opiates, amphetamines, and nicotine; plasma testing for cocaine and alcohol; and hair analysis for cocaine. Enrolled women will be followed by telephone to ascertain information about pregnancy complications and drug use. Any participants experiencing spontaneous abortions at HUP through 22 weeks of pregnancy will have pathologic and chromosomal analysis of the aborted fetus/embryo and placenta. Univariate analyses, including calculation of odds ratios with confidence intervals, will be followed by stratification and multiple logistic regression, enabling us to evaluate the importance of cocaine on spontaneous abortion while adjusting for other potential confounders. The prevalence of pathologic and chromosomal changes in spontaneous abortuses from mothers using and not using cocaine will also be determined. This research will provide an enhanced understanding of the impact of cocaine use, a potentially modifiable risk factor, on early fetal mortality. Such a study is crucial to any successful effort for understanding and preventing abnormalities in the offspring of cocaine-using mothers.